Shared signaling pathways among gasotransmitters.

G asotransmitters, including NO, CO, and H2S, are a group of endogenously produced gas molecules that are membranepermeable and share similar molecular targets (1). Gasotransmitters may antagonize or potentiate each other’s cellular effects at three levels: their production, their downstream molecular targets, and the direct chemical interaction among themselves. The study published in PNAS by Coletta et al. (2) presents an interesting example of how the interaction of gasotransmitters converges at the same downstream molecular target. Coletta et al. found that the proangiogenic effect of H2S on angiogenesis and wound healing was completely absent in endothelial NO synthase KO mice, and that eliminating H2S production by silencing cystathionine-γ-lyase (CSE) abolished NO-stimulated angiogenesis (2). This mutually dependent relationship between H2S and NO in vascular endothelial cells is rooted in the regulation of the cellular levels of cGMP. NO activates soluble guanylyl cyclase (sGC) to generate cGMP, whereas H2S inhibits phosphodiesterase-5 (PDE5) to slow down the degradation of the existing cGMP (2). A net increase in cGMP level leads to the activation of protein kinase G (PKG) and its downstream effector, vasodilator-stimulated phosphoprotein. Lack of NO might curtail the production of cGMP so that H2Sinhibited cGMP degradation becomes meaningless. Conversely, rapid degradation of the existing cGMP in the absence of H2S would render NO powerless to elicit angiogenesis. The activation of sGC by NO and CO has been well documented (3). The affinity of NO to sGC is approximately 30 to 100 times greater than that of CO (4). Hitherto, there has been no evidence that H2S actually interacts with sGC. It had been suggested, however, that endogenous H2S might increase cGMP level by inhibiting PDE activity in the heart (5). This suggestion was later verified by Bucci et al. with cultured rat aortic smooth muscle cells (SMCs) (6). It is worth pointing out that H2S appears to inhibit different subtypes of PDE without specificity. As such, H2S may inhibit the degradation of both cAMP and cGMP in different types of cells. Depending on whether it is cAMP or cGMP level that is being affected, the cellular effects of H2S may be quite different. Compared with their effects on angiogenesis, individual and combined effects of NO and H2S on endothelium-dependent vasorelaxation are more complex. After being synthesized in, and released from, the endothelium, NO activates sGC in vascular SMCs to induce vasorelaxation (3), whereas H2S opens intermediateand small-conductance calcium-sensitive K channels in the endothelium as an endothelium-derived hyperpolarizing factor (7), and stimulates KATP channels in vascular SMCs to relax blood vessels (8). A mechanism was introduced by Coletta et al., who ascribed endothelium-dependent and H2S-induced vasorelaxation partially to the inhibition of PDE5 in vascular SMCs (2). This proposed mechanism is not undisputed, as, under resting conditions, total cGMP content in rat aortic rings was not significantly decreased by silencing CSE, or increased by addition of NaHS (a H2S donor) (2). Another important consideration for NOor H2S-induced endotheliumdependent vasorelaxation is the roles played by individual gasotransmitters in different types of blood vessels. It has been proposed that NO may play a more important vasorelaxant role in conduit artery (e.g., aorta) as an endotheliumderived relaxing factor but that H2S may be more critical in regulating the relaxation of peripheral resistance arteries (e.g., mesenteric arteries and coronary arteries) as an endothelium-derived hyperpolarizing factor (7, 9). Previous studies have shown that H2S is approximately five times more potent in relaxing the rat mesenteric artery vs. the rat aortic artery (9). In CSE-KO mice, the methacholine-induced endothelium-dependent vasorelaxation was significantly decreased (10), which may or may not be related to changes in cGMP level, as the blood vessel under investigation is a classical resistance mesenteric artery (10), whereas the proposed inhibition of PDE5 by H2S was derived from studies on rat aortic tissues (2). Whether H2S-induced vasorelaxation is mediated by the cGMP pathway, wholly or partially, is still unsettled. NO-induced vasorelaxation was abolished by ODQ or NS2028. However, these two sGC inhibitors neither had effect on, nor potentiated, H2S-induced relaxation of rat aortic tissues (11). Also tested on rat aortic tissues, NaHS-induced relaxation was mostly inhibited by the blockade of KATP channels and to a much lesser extent by ODQ (12). The vasorelaxant effect of NaHS on rat coronary arteries is also unaffected by Nω-nitro-L-arginine methyl ester or ODQ (13). The precontracted chicken ductus arterious rings were relaxed by Na2S (another H2S donor) or by authentic CO. The vasorelaxant effect of CO is inhibited by Nω-nitro-L-arginine methyl ester or ODQ, but that of Na2S is not affected (14). Differences in these vasoactive studies regarding the mediation of H2S effect by cGMP level cannot yet be readily explained. At any rate, it is too early to conclude that the vasorelaxant effects of H2S and NO converge at the cGMP-PKG node. Alterations in KATP channel openings, intracellular free calcium levels, oxidative stress status, and other signaling events may all be involved in the vasoactive effects of H2S, irrespective of changes in cGMP level in different vascular preparations or in the same type of blood vessels from different species. Cellular effects of gasotransmitters, and the signaling mechanisms for the same, are cell type-specific. Endothelial cells and SMCs make this case in point. We can start with the production of gasotransmitters. NO production is virtually a patented function of eNOS in endothelial cells. H2S, however, can be produced in both endothelial cells and SMCs. As such, the direct effect of endogenous H2S on SMCs would unlikely be affected by endogenous NO in the absence of the endothelial cell. Coletta Fig. 1. Signaling convergence of gasotransmitters in vascular SMCs. ER, endoplasmic reticulum; IKCa/ SKCa, intermediateand small-conductance calciumsensitive K channel; KATP, ATP-dependent K channel.